BayesianFitForecast

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #ai #typescript #llm #go #golang #android #machine-learning #rust #deep-learning #linux

BayesianFitForecast

★ 10 Open GitHub ↗

BayesianFitForecast: A comprehensive R toolbox for Bayesian parameter estimation and forecasting in ODE models, making advanced Bayesian methods accessible across diverse scientific domains.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# BayesianFitForecast

**BayesianFitForecast** is an R toolbox designed to streamline Bayesian parameter estimation and forecasting in ordinary differential equation (ODE) models, particularly for infectious disease applications. This toolbox simplifies the complex process of Bayesian inference by automating the generation of Stan files, enabling users to configure models, define priors, and analyze results efficiently, even with minimal programming expertise.

<p> <a href="https://arxiv.org/abs/2411.05371" target="_blank">BayesianFitForecast Tutorial</a></p> 
<p><a href="https://www.youtube.com/watch?v=jnxMjz3V3n8" target="_blank">Video Tutorial</a></p>

## Features

- **Automated Stan File Generation**: Automatically generate Stan files based on user-defined ODE models, eliminating the need to write code manually.
- **Bayesian Parameter Estimation**: Estimate model parameters using Bayesian inference, incorporating uncertainty and prior knowledge, which is particularly valuable in contexts with limited or noisy data.
- **Model Performance Evaluation**: Includes comprehensive tools for evaluating model performance, such as convergence diagnostics, posterior distributions, credible intervals, and advanced performance metrics (e.g., Mean Absolute Error, Mean Squared Error, and Weighted Interval Score).
- **Versatility**: Applicable to a wide range of scientific fields, with demonstrated use cases including the analysis of the 1918 influenza pandemic in San Francisco using different error structures (e.g., Poisson and negative binomial) within the SEIR model.
- **Tutorial Resources**: Includes a detailed tutorial and video guide to help users get started with the toolbox.

## Usage

### 1. Configuring Your Model

Start by defining your ODE model, parameters, and priors in the `options_generaluse.R` file. This file allows you to specify the model variables, parameters, error structures, and priors.

### 2. Running the Model

After configuring your model, use the `run_MCMC.R` script to fit the model to your data. This script performs the Bayesian parameter estimation using the Stan files automatically generated by the toolbox.

### 3. Analyzing Results

Once the model fitting is complete, use the `run_analyzeResults.R` script to generate and analyze the results. This script will produce various outputs, including Excel files with the parameter estimates, performance metrics, and convergence diagnostics.

### 4. Example: SEIR Model for the 1918 Influenza Pandemic

This toolbox includes several predefined case studies, such as the SEIR model used to analyze the **1918 Influenza Pandemic** in San Francisco. The toolbox allows users to compare different error structures (e.g., Poisson vs. negative binomial) and assess the quality of model fit with performance metrics such as Mean Absolute Error (MAE) and Weighted Interval Score (WIS).

## License

This project is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0). See the LICENSE file for details.

## Citation

If you use this toolbox in your research, please cite the following paper:

Karami, H., Bleichrodt, A., Luo, R. et al. BayesianFitForecast: a user-friendly R toolbox for parameter estimation and forecasting with ordinary differential equations. BMC Med Inform Decis Mak 25, 385 (2025).[LINK](https://doi.org/10.1186/s12911-025-03208-z)

BayesianFitForecast

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

BayesianFitForecast

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README